An Incremental Unified Framework for Small Defect Inspection (2023-12-14T00:00:00.000000Z)

TL;DR

The Incremental Unified Framework (IUF) is presented, which can reduce the feature conflict problem when continuously integrating new objects in the pipeline, making it advantageous in object-incremental learning scenarios.

Abstract

Artificial Intelligence (AI)-driven defect inspection is pivotal in industrial manufacturing. Yet, many methods, tailored to specific pipelines, grapple with diverse product portfolios and evolving processes. Addressing this, we present the Incremental Unified Framework (IUF), which can reduce the feature conflict problem when continuously integrating new objects in the pipeline, making it advantageous in object-incremental learning scenarios. Employing a state-of-the-art transformer, we introduce Object-Aware Self-Attention (OASA) to delineate distinct semantic boundaries. Semantic Compression Loss (SCL) is integrated to optimize non-primary semantic space, enhancing network adaptability for novel objects. Additionally, we prioritize retaining the features of established objects during weight updates. Demonstrating prowess in both image and pixel-level defect inspection, our approach achieves state-of-the-art performance, proving indispensable for dynamic and scalable industrial inspections. Our code will be released at https://github.com/jqtangust/IUF.

Authors

Jiaqi Tang

4 papers

Ruizheng Wu

3 papers

Xiaogang Xu

3 papers

TL;DR

Abstract

Authors

References45 items

Unsupervised Continual Anomaly Detection with Contrastively-learned Prompt

Defect Spectrum: A Granular Look of Large-Scale Defect Datasets with Rich Semantics

OmniAL: A Unified CNN Framework for Unsupervised Anomaly Localization

Detecting Anomalous Solder Joints in Multi-sliced PCB X-ray Images: A Deep Learning Based Approach

SimpleNet: A Simple Network for Image Anomaly Detection and Localization

SVDiff: Compact Parameter Space for Diffusion Fine-Tuning

A Comprehensive Survey of Continual Learning: Theory, Method and Application

Pushing the Limits of Fewshot Anomaly Detection in Industry Vision: Graphcore

Tire Bubble Defect Detection Using Incremental Learning

Automated Continual Learning of Defect Identification in Coherent Diffraction Imaging

Towards Continual Adaptation in Industrial Anomaly Detection

A New Knowledge Distillation Network for Incremental Few-Shot Surface Defect Detection

SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentation

A Unified Model for Multi-class Anomaly Detection

Continual Learning with Lifelong Vision Transformer

A continual learning framework for adaptive defect classification and inspection

Deep-learning-based anomaly detection for lace defect inspection employing videos in production line

UTRAD: Anomaly detection and localization with U-Transformer

Efficient anomaly detection from medical signals and images with convolutional neural networks for Internet of medical things (IoMT) systems

DRÆM – A discriminatively trained reconstruction embedding for surface anomaly detection

Siamese Neural Network Based Few-Shot Learning for Anomaly Detection in Industrial Cyber-Physical Systems

Defect Detection Using Deep Lifelong Learning

Towards Total Recall in Industrial Anomaly Detection

Inpainting Transformer for Anomaly Detection

CutPaste: Self-Supervised Learning for Anomaly Detection and Localization

Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks

Cross Fusion Net: A Fast Semantic Segmentation Network for Small-Scale Semantic Information Capturing in Aerial Scenes

PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization

PANDA: Adapting Pretrained Features for Anomaly Detection and Segmentation

ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting

Sub-Image Anomaly Detection with Deep Pyramid Correspondences

HRank: Filter Pruning Using High-Rank Feature Map

Towards Visually Explaining Variational Autoencoders

MVTec AD — A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection

Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders

GANomaly: Semi-Supervised Anomaly Detection via Adversarial Training

Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence

Memory Aware Synapses: Learning what (not) to forget

Gradient Episodic Memory for Continual Learning

Continual Learning Through Synaptic Intelligence

Overcoming catastrophic forgetting in neural networks

Anomaly detection for industrial surface inspection: application in maintenance of aircraft components

Federated learning-based Explainable Anomaly Detection for Industrial Control Systems

Analysis and design of an adaptive minimum reasonable inventory control system

An Incremental Unified Framework for Small Defect Inspection 17

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names